Add b-jet preselection for the jet trigger

This starts on fast b-tagging preselection to the jet trigger.

The end result is that we get three flavor tagging probabilities (b, c, and light, called fastDips_p{c,u,b}) on each of the EMTopo trigger jets we use for preselection before running fullscan tracking.

We made some changes to flavor tagging, and jet trigger code while we were at it:

• The offline jet configuration code has a tradition of passing around container names without the Jets suffix, and then adding the Jets in every function that actually references a jet collection. We'd followed this custom in the b-tagging group, but it runs into problems with HLT jet collections (which we're trying to b-tag here) because they have no Jets suffix. Rather than try to make each configuration function clever enough to insert / remove Jets in the right places in container names, we're opting to explicitly add the Jets suffix where needed (or, in some cases, to not remove it).
• We did a bit of cleanup for ATLASRECTS-6635, removing calls to CompFactory on import where we could.
• On any files we touched we cleaned up line endings and trailing whitespace.
• We we moved import statements to the top of files where we could.

Most development was tested with

test_trig_mc_v1Dev_slice_jet_build.py

and

test_trig_data_v1Dev_slice_jet_build.py

which seemed to run fine.

Open issues

• The Trigger_athenaHLT.citest and Trigger_athena_data.citest fail because of missing beamspot covariance information.
• TriggerMenuMT_generateMenuMT_newJO_ctest fails because NameError: name 'include' is not defined. I'm wondering how include managed to get into "newJO".